SAFE
Millwrights spend their hours rigging and setting multi-ton machinery, laser-aligning shafts to thousandths, shimming baseplates, dialing in couplings and gearboxes, and cutting/welding in live plants — work that happens in cramped, dirty, non-standardized spaces no current robot can navigate. AI can absorb the paperwork tier: reading prints, generating alignment and torque calculations, interpreting vibration spectra, writing PM schedules and repair reports. What it cannot do is feel a bearing preload, decide mid-shutdown whether to rebuild or replace with the line down at $50k/hour, or own that call.
This fall is concentrated in 2020 and has not recovered since.
Median pay $55,560 → $65,700 -5.4% in real terms
This line is counted by the Bureau of Labor Statistics — the one figure on this page that isn't a judgement of ours. Headcount moves on demand, offshoring, demographics and the business cycle, and automation is one term among several, often not the loudest.
So a falling line is not evidence that AI did it, and a rising one is not evidence that it won't. Both happen in this register: some occupations resist automation and shrink anyway, others are highly automatable and keep growing. The marked year is 2020.
BLS projection, 2024–2034
0% 41,300 → 41,300 on the projections basis
Hard to automate, but shrinking anyway
The work resists current AI, yet the BLS projects 0% by 2034. Whatever is shrinking this occupation, the evidence does not point to automation — demand, demographics, offshoring and industry decline all shrink jobs that no machine could do.
Different clocks. The score is what current AI could do to this work today. The projection is how many of these jobs will exist in 2034. Everything between the two — how fast employers actually adopt, whether demand grows in the meantime — is why they can point opposite ways without either being wrong.
~3,600 openings a year on average, including replacing people who leave.
MoverMechanicMillwrightGear RepairerMachine MoverPlant ChangerMachine RiggerMachine SetterMachine ErectorMachinery MoverMachinery RiggerMachine InstallerMachinery ErectorMillwright HelperMillwright ForemanMillwright MechanicMachinery DismantlerMaintenance MechanicPrecision MillwrightIndustrial MillwrightJourneyman MillwrightMillwright ApprenticeMillwright InstructorMaintenance Millwright
Holding it up: embodiment . Weakest point: liability shield .
Tasks largely resist digitisation The bulk of a millwright's shift — chain-falling a 4,000 lb gearbox into position, dial-indicating a shaft to 0.002", grinding a shim pack to fit, tacking a base to embedded plate — has no digital substitute, and the 17 rather than 20 reflects that the calculation, print-reading, precision-laser math and vibration-spectrum interpretation layers are already being handed off to software and predictive-maintenance platforms.
Hands-on in uncontrolled environments You work inside the machine: crawling into conveyor housings, hanging off scaffold in a boiler house, hot work in confined space with a permit and a fire watch, temperatures and dust and noise that vary shop to shop and shutdown to shutdown — there is no controlled-environment version of this task, which is why it sits at the ceiling.
Certification preferred, not legally required There is no millwright licence — the credentials in play are NCCER, journeyman card from a UBC or IAM apprenticeship, rigging cert, hot-work and confined-space permits, and OSHA 10/30, all employer- or site-required rather than state-issued, so the liability for a dropped load or a failed weld runs to the contractor and the plant, not to your name on a stamp.
Exists to be accountable for ambiguous calls With the line down and the clock running you decide whether that gearbox gets rebuilt in place or swapped, whether soft-foot is bad enough to justify pulling the baseplate, whether the crane pick is safe with the rigging on hand — calls made on incomplete information with real cost and injury exposure, though a plant engineer or maintenance supervisor typically signs off on the big-ticket replace decision, which keeps it at 14 rather than higher.
The verdict above describes this occupation as a whole. Almost nobody does the typical version of a job — tick what's actually in your week and see how your own mix sits.
Your task mix speaks to task resistance (17/20 here) — how much of the day's work current AI already does. That is the dimension the boxes above are about.
It cannot move the other three. Liability shield (6/20) is whether the law requires a licensed human to sign. Trust premium (11/20) is whether buyers specifically pay for a person. Judgment and accountability (14/20) is whether the role exists to own consequential calls. Those are facts about the occupation's standing, not about which tasks are in your week — a paralegal who does only trial exhibits still holds no licence. Together they are 31 of this occupation's 68 points (46%).
Embodiment (20/20) is also a property of the work rather than the worker, but we don't tag individual tasks as physical or not, so the picker can't tell you anything about it. That's a limit of this tool, not a claim.
Did we get the list right? Tell us what's missing — the tasks are written from the outside, and you're reading this from the inside.
The moves above are yours to make. This is the other half: what would have to change in the world for the occupation itself to score higher. None of it is in any one person's gift, but it is where the floor actually comes from. Scores here are not a one-way ratchet. Only two of the five dimensions — task resistance and embodiment — track what machines can do. The other three track law, what buyers will pay for, and who is answerable, and those move in both directions, often in response to the same pressure AI creates. If every lever below landed, this occupation would score around 83/100, still SAFE.
ASME/API machinery-alignment and boiler/pressure-equipment reinstallation records requiring a certified technician's stamp — e.g. National Board R-stamp welding repair documentation already requires certified welders; broader adoption of similar traceable sign-off for rotating-equipment reinstall in NFPA 70E/insurance-audited plants
Task-mix shift: as vibration-spectrum interpretation, torque math and PM scheduling get absorbed by software, the remaining day is dominated by rebuild-vs-replace calls during live shutdowns at $50k/hour, which the millwright already owns in practice; formalizing that as shutdown-lead authority in maintenance contracts or union scope language raises it further
Rigging and critical-lift sign-off becoming a named, personally-accountable role: OSHA 1926.1400-style qualified-rigger/lift-director certification already exists in construction cranes, and extension of a written critical-lift-plan-with-signature requirement to general industry machinery installation (or insurer mandates from FM Global / Zurich for lifts above a tonnage threshold) would put a named millwright's signature and liability on setting work
Turnaround/shutdown contracting where the plant specifically buys a named crew with a documented alignment track record rather than lowest-bid labor — visible in petrochemical turnaround contracts and OEM-authorized installer programs (Sulzer, Flowserve, SKF certified) that condition warranty on a certified installer
Precision alignment tolerances tightening as high-speed and larger-frame equipment spreads (turbomachinery, large gearboxes in wind and battery plants), pushing more of the day into hands-on shimming, soft-foot correction and thermal-growth compensation that no software substitutes for
The limit. Already 68 with embodiment maxed; the realistic headroom is liability and named-signature structure, not capability. Millwright work has no licensing board equivalent to electricians or PEs, so a full statutory shield is the least likely path — the credible route is insurer and OEM-warranty requirements rather than state law.
| Chicago-Naperville-Elgin, IL-IN | 1,490 | $74,730 +14% |
| Detroit-Warren-Dearborn, MI | 1,490 | $91,850 +40% |
| Houston-Pasadena-The Woodlands, TX | 820 | $67,750 +3% |
| Los Angeles-Long Beach-Anaheim, CA | 710 | $99,250 +51% |
| Dallas-Fort Worth-Arlington, TX | 640 | $65,500 +0% |
| Kansas City, MO-KS | 640 | $91,380 +39% |
| Atlanta-Sandy Springs-Roswell, GA | 510 | $64,160 -2% |
| Philadelphia-Camden-Wilmington, PA-NJ-DE-MD | 510 | $77,120 +17% |
| San Francisco-Oakland-Fremont, CA | 320 | $113,490 +73% |
| Kennewick-Richland, WA | 80 | $107,700 +64% |
| Los Angeles-Long Beach-Anaheim, CA | 710 | $99,250 +51% |
We have no reported case of a named organisation automating this occupation. Not one deployment, not one announcement.
That is worth saying out loud next to a score of 68. The verdict above is about what the work exposes — what current AI could do to these tasks. It is not a claim that anyone has done it. For this occupation those two things have come apart completely: the capability argument is on this page, and the evidence column is empty.
Has AI actually changed your work? One tap, anonymous, and the running tally is public. Nothing else is asked of you.
Rather than check back: get the digest and we'll tell you what changed — or watch a single occupation from its own page.